Fluoroscopic Analysis of Lumbar Epidural Contrast Spread After Lumbar Interlaminar Injection

Methods 

One hundred fourteen consecutive patients undergoing interlaminar lumbar ESI were prospectively studied. We used a 20-gauge Tuohy needle for all injections and the loss-of-resistance technique with the patient in the prone position. All injections were performed under fluoroscopy. Proper epidural placement was confirmed, both with loss of resistance and with fluoroscopy, using anteroposterior (AP) and lateral views to check needle placement. All injectionists were experienced and all had completed interventional pain management fellowships. During the injection procedure, fluoroscopic images were taken to determine contrast spread at a specific injectate volume of 5mL in both the AP and lateral views for the purpose of standardization. The fluoroscopic images were then evaluated for various criteria regarding the spread of contrast, discussed below. The contrast media was iohexol (Omnipaque 300).a Approximately 2mL of iohexol and 3mL of medication (80mg of triamcinolone acetonide, 1mL of preservative-free 1% lidocaine, and 5mL of preservative-free normal saline) were injected and the fluoroscopic image was then taken. A standard volume of 5mL was used when the fluoroscopic images were taken. We used the OEC 9800 C-arm fluoroscopeb with all subjects.

We evaluated 5 parameters for contrast spread. The first was whether the epidural spread was unilateral (left or right only) or bilateral. The second was whether needle tip placement was within the confines of the lateral border of the spinous process distal end, which counted as within or medial (fig 1). We defined lateral to the borders of the distal spinous process as lateral for the purpose of needle placement (fig 2). The third parameter was whether there was an anterior (ventral) epidural spread. Because there was posterior spread in all cases, anterior spread in combination with posterior spread was defined simply as anterior spread. The fourth parameter was whether the contrast spread caudally, at least 1 lumbar segment down on the lateral view, and the fifth parameter was the patients’ sex and the level of injection. The other variables were compared against the patients’ sex and level of injection, which was always at L4-5 or L5-S1. Other levels were excluded to ensure standardization. The first author (LW) interpreted the fluoroscopic images and the second author (NHF) performed a repeat, blinded interpretation. All disagreements between the 2 interpretations were discarded (a total of 4 cases).

View Large Image Download to PowerPoint

Fig 1. Needle tip medial to the border of the spinous process.

View Large Image Download to PowerPoint

Fig 2. Needle tip lateral to the border of the spinous process.

Results 

We studied 114 patients (63 men, 51 women). The needle placement was within the boundaries of the spinous process 53% of the time and was lateral to or outside the borders of the spinous process 47% of the time.

Contrast spread was anterior in 47% of the patients and remained posterior in 53% of the patients. The spread of the contrast and medication to more than 1 level caudally occurred 76% of the time and to less than 1 level caudally only 24% of the time.

With regard to needle position, when the needle tip was placed lateral to the border of the spinous processes, the spread was unilateral 76% of the time. Bilateral spread occurred in only 24% of cases when the needle was outside of the lateral border.

When the needle tip was medial to the border of the spinous process at final placement (close to midline), the spread was unilateral 45% of the time and bilateral 55% of the time. Whether the needle tip was medial or lateral, the contrast spread more than 1 interspace caudally 76% of the time.

With the needle tip medial to the border of the distal end of the spinous process, the contrast spread anteriorly 51% of the time and remained only posterior 49% of the time. With lateral needle tip placement, contrast spread anteriorly 43% of the time and remained posterior 57% of the time.

The sex of the patients did not have any appreciable effect on anterior and posterior spread. Spread of more than 1 segment caudally occurred 71% of the time in men and 83% of the time in women.

Table 1 summarizes the effect of the location of needle tip placement on contrast spread. Table 2, Table 3 summarize the effects of the level of site selected with reference to needle tip position on the spread of the contrast.

Table 1. Effect of Medial and Lateral Needle Tip Placement on Contrast Spread

	Location of Needle Tip Placement	% of Sample	% Contrast Spread Unilateral/Bilateral	% Contrast Spread Anterior/Posterior
	Medial to the border of the spinous process	53	45/55	51/49
	Lateral to the border of the spinous process	47	76/24	43/57

Table 2. Effect of Lumbar Level Selection on Anterior and Posterior Contrast Spread

	Level of Site Selected	Total % Contrast Spread Anterior/Posterior	% Contrast Spread Anterior/Posterior With Needle Tip Placement Medial to the Border of the Spinous Process	% Contrast Spread Anterior/Posterior With Needle Tip Placement Lateral to the Border of the Spinous Process
	L4-5	50/50	47/53	37/63
	L5-S1	46/54	46/54	45/55

Table 3. Effect of Lumbar Level Selection on Unilateral and Bilateral Contrast Spread

	Level of Site Selection	% Contrast Spread Unilateral/Bilateral With Needle Tip Placement Medial to the Border of the Spinous Process	% Contrast Spread Unilateral/Bilateral With Needle Tip Placement Lateral to the Border of the Spinous Process
	L4-5	39/61	81/19
	L5-S1	50/50	74/26

Unilateral spread versus bilateral spread appeared to have a mild effect on whether the contrast reached anteriorly into the epidural space. Unilateral spread with anterior spread occurred 45% of the time and unilateral with posterior-only spread occurred 55% of the time. Bilateral epidural spread with anterior spread occurred 53% of the time, and bilateral, posterior- only occurred 47% of the time.

Discussion 

In this study, we found that the contrast generally spreads caudally more than 1 level in the majority of cases. Because we excluded postsurgical patients from the study, there were no adhesions to distort its result. Overall, the contrast spread caudally more than 1 full vertebral level 76% of the time. This occurred slightly more often in women (82% of the time) than in men (71% of the time), most likely because women are smaller and the volume of injectate (5mL) was kept constant for all patients. We found that the spread of the contrast caudally more than 1 level was not affected significantly by the medial or lateral placement of the needle tip, or by precisely at what level it was injected. In the past, many physicians have assumed that a craniad direction of the injectate generally takes place without any significant caudad spread. This study shows that caudad spread of at least 1 level or more does occur in slightly more than 75% of the injections.

Unilateral versus bilateral spread of contrast and medication appears to be affected by needle placement. If the needle is placed close to the midline, or within the lateral borders of the spinous process, bilateral spread occurs more than 55% of the time. If the needle tip is placed lateral to the borders of the spinous process, however, unilateral spread occurs 76% of the time. Therefore, for bilateral spread, a needle tip placed close to the midline is likely optimal. One issue with this study was determining exact midline placement, which would, in our opinion, be open to some bias. A close-to-midline placement of the needle tip was estimated as being within the lateral borders of the distal end of the spinous process immediately above the needle placement. We suggest, however, that as the needle approached even closer to the midline, the spread more often would be bilateral. The reason for the unilateral spread if the needle is off midline is somewhat vague. A thin membrane, the plica mediana dorsalis, that separates the 2 halves of the epidural space on the AP view has been documented.8 This likely accounts for the unilateral spread that is seen more than 75% of the time with lateral needle placement.

Thus, if one desires an optimal bilateral spread, an interlaminar ESI at the L5-S1 level should be performed under fluoroscopy, with careful attention given to having a midline location at insertion of the needle tip. This also holds true at the L4-5 level although to a slightly lesser extent.

Two studies previously addressed the spread of contrast in lumbar epidurals. The first was a detailed retrospective review by Whitlock et al.1 The second, by Botwin et al,2 was a prospective study with a limited number of patients. The studies had different results. Whitlock indicated that anterior epidural flow occurred in 79% of all cases. The second study however, found anterior epidural flow in only 36% of the cases. (Our study revealed anterior epidural flow in 47% of the cases.) With regard to bilateral spread, Whitlock found that 88% of all epidural injections reviewed spread bilaterally, with 92% bilateral spread if the needle was close to midline (within spinous processes). Botwin’s limited prospective study found that only 16% of the cases had bilateral spread. As previously noted, however, Botwin did not account for needle tip placement. Given that the spread in the Botwin study was unilateral 84% of the time, we assume the needle was lateral to the borders of the spinous processes in almost every case.

Needle tip placement distribution in Whitlock et al1 was roughly similar to ours in terms of medial versus lateral placement. Results differed, however. When our needle was placed medially, the contrast spread bilaterally 55% of the time (vs 92% in the Whitlock study). When our needle was placed laterally, bilateral spread occurred only 24% of the time (vs 83% in Whitlock’s study).

There could be multiple explanations for the discrepancies in these results. The first relates to methodology. Whitlock stated that the needle was often repositioned to ensure bilateral spread of the medication as a routine during epidural injections. It appears this was the protocol at their institution. Clearly, this would lead to a bias toward bilateral spread of the injectate.

The second explanation involves interpretation of the fluoroscopic images. Whitlock had a medical student read the images, with a radiologist double-checking approximately 1 in 10 of the images. In our study, the first author read the images, with a blinded double check performed by another author. All coauthors of this study are board-certified pain physicians. If there was disagreement, the case was discarded (4 cases were discarded). During our study, we noted that it could be difficult to properly read fluoroscopic pictures retrospectively, especially with regard to anterior spread versus artifacts. Often, artifacts or bony shadows can appear as anterior spread. A retrospective reading by a less experienced professional can often result in over-reporting anterior spread.

A third explanation is that there was a difference in volume of injectate between the Whitlock study (12mL) and ours (5mL). A higher volume may certainly affect the spread in the epidural space, although further study would be needed to confirm this.

The key findings in our study are that contrast spreads in a caudad direction (1 segment or more) approximately 75% of the time, and that needle placement greatly affects whether contrast and medicines spread unilaterally or bilaterally. Bilateral spread is seen more often with midline placement. Again, because we allowed midline placement essentially to be defined as within the borders of the lamina above, a relatively large area was defined as midline. Anterior or posterior spread did not appear to be significantly affected by midline or lateral placement.

The tendency to spread caudally as well as cranially also does not appear to be significantly affected by needle placement. Presumably, spread of contrast and medicine over the area of pathology would correlate with pain relief. This was shown by Winnie et al6 with regard to level. It is reasonable to assume that spread would be very important when it is unilateral, particularly if there is radicular pain on the opposite side. This would again argue for the use of fluoroscopy for epidural placement.

Last, we found anterior spread of contrast to be present about 50% of the time on epidural injection at a volume of 5mL of injectate. It is unknown whether anterior spread of injectate directly relates to the efficacy of the injection for pain relief. Further investigation into whether anterior spread of epidural injectate correlates with pain relief would be interesting. It is clear, however, that anterior spread can frequently be achieved with interlaminar injections.

Conclusions 

With other variables kept constant, needle placement—specifically, whether close to midline or more lateral from midline—affects unilateral versus bilateral spread of contrast and medicine. Needle placement medial to the border of the spinous process significantly increases the frequency of bilateral spread and mildly increases the frequency of anterior spread. Contrast spreads more than 1 level caudally more than 75% of the time regardless of the location of needle placement within the epidural space. Women had a slightly higher percentage (82%) spread of contrast greater 1 level caudally than did men (71%).

Suppliers